Every SEO I know has an opinion on AI Overviews using their content without sending traffic back. Fair enough. But a lot of those same sites are doing the exact same thing to smaller creators they quote every week.

Publishers are angry that AI Overviews pull from their content without sending much traffic back. LeRoy doesn't dismiss that frustration. He just points out that plenty of publishers do the exact same thing to smaller sites.

His example: SEOJobs.com, a site he runs, had its data used by a major publisher. They named the site in the article. They didn't link to it.

That's the pattern. A source gets mentioned by name, sometimes even spelled out as plain text, but never made clickable. The citation shows up as a courtesy instead of a standard.

Link equity is the idea that a page passes some of its ranking authority through the links it makes. Plenty of sites treat it like water in a bucket: link out, and you're pouring some of your own strength away.

LeRoy pushes back on that directly. Linking to a relevant, high-quality source doesn't drain your rankings. If you're avoiding citations because you think you're protecting your site, you're working from a theory that doesn't hold up.

He also shares a stranger example. He once asked a publisher to link to a dataset they were already referencing from his site. They said no, then offered to write a paid post about his site for $500 instead.

That's not an editorial decision. That's gatekeeping dressed up as policy, and it turns a research citation into a sales opportunity.

Even when sites do link out, they often default to nofollow. Google actually offers three distinct attributes here: sponsored for paid placements, ugc for user-generated content like comments, and nofollow for when you genuinely don't want to be associated with a destination.

Slapping a blanket nofollow on every outbound link, including ones you trust enough to send your own readers to, undercuts the whole point of citing someone. It's the same complaint people have about AI, aimed downward: use the work, share almost none of the benefit back to whoever made it possible.

That's really the core of it. The complaint about AI is that someone else uses your work to improve their product and you get nothing back. Publishers do this to smaller researchers and creators constantly, strengthening their own articles with someone else's data while denying that person a single visit.

LeRoy proposes a simple checklist before you hit publish on anything that references someone else's work:

If the answer to all four is yes, give it a normal, clickable, followed link. No link-juice theory required.

Pull up your last five published pieces and check how you handled sources. Named without a link, quietly nofollowed, or actually credited the way you'd want to be credited?

If the pattern looks more like hoarding than citing, that's worth fixing before writing another word about what AI owes you.

Did You Know? Google splits the rel attribute for outbound links into three separate categories: sponsored, ugc, and nofollow. Most sites still apply a single blanket nofollow regardless of which one actually fits.

One thing I'm thinking about: the open web only works if it's actually webbed together. We spend a lot of energy worried about AI platforms scraping our work and keeping the value for themselves, and that worry is legitimate.

But LeRoy's point lands because it's uncomfortable. We can't play link police with Google while acting as link hoarders on our own sites.

Next time you reference a study or a piece of original reporting, ask whether you're giving that source the same transparency you're demanding from the AI giants. It's a simple standard. We just have to actually follow it.

Till next time,

Source